Machine Learning in Linux: FBCNN – JPEG artifacts removal

Posted by sde on Mar 28, 2023 2:44 AM EDT
LinuxLinks.com; By Steve Emms
Mail this story
Print this story

FBCNN (flexible blind convolutional neural network) is software which seeks to remove artifacts from JPEGs while preserving the integrity of the images. It decouples the quality factor from the JPEG image via a decoupler module and then embeds the predicted quality factor into the subsequent reconstructor module through a quality factor attention block for flexible control.

Full Story

  Nav
» Read more about: Story Type: Reviews, Tutorial; Groups: Linux, Python

« Return to the newswire homepage

This topic does not have any threads posted yet!

You cannot post until you login.